Why Modeling Cultural Evolution Is Still Such a
Challenge
Dan Sperber and Nicolas Claidière
Institut Jean-Nicod (CNRS, EHESS, ENS), Paris, France
Correspondence to: dan@sperber.com
The idea that cultural evolution exhibits variation, competi-
tion, and inheritance and therefore can be studied by adjusting
the Darwinian theory of evolution by natural selection is an
attractive one. It has been argued by a number of authors (e.g.,
Campbell 1960; Monod 1970; Dawkins 1976; Cavalli-Sforza
and Feldman 1981; Boyd and Richerson 1985; Durham 1991;
Aunger 2002; Mesoudi et al. 2004) and pursued in a variety
of ways, some (Dawkins and memeticists) staying close to
the Darwinian model, others (e.g., Boyd, Richerson, and their
collaborators) being more innovative. We agree that there are
relevant analogies between biological and cultural evolution
and, in particular, that cultural items do exhibit variation, com-
petition, and cumulative modification. On the other hand, we
believe that a proper understanding of the mechanisms of cul-
tural propagation drawing on the work of cognitive and social
scientists (see Sperber and Hirschfeld 1999 for a review) con-
tradicts the idea that culture exhibits inheritance in the strict
sense needed for the theory of evolution by natural selection
to apply straightforwardly to it. If so, it will take more than
adjusting the Darwinian model to be faithful to the Darwinian
inspiration.
Peculiarities of Cultural Propagation
Whereas the mechanisms of biological inheritance operate at a
molecular level, invisible to the naked eye, and their discovery,
from Mendel to Crick and Watson and beyond, has taken the
form of a series of major scientific advances, the mechanisms
of cultural inheritance are, to a significant extent, accessi-
ble to ordinary observation. Cultural propagation—learning,
teaching, sharing of attitudes and values, and so on—takes
place through the ordinary channels of information trans-
mission, and in particular through imitation and communi-
cation. Imitation, communication, and also memory (without
which cultural information would not survive to propagate)
20 Biological Theory 1(1) 2006, 20–22.

have been studied in depth by neuroscientists, psychologists,
linguists, anthropologists, and sociologists. Drawing on their
work, one can state with confidence that cultural propagation
differs from biological propagation in crucial respects.
Among the various mechanisms involved in biologi-
cal heritability (cytoplasmic heritability, sex linkage dise-
quilibrium, epigenetic transmission, niche construction, and
so on), replication of DNA is the most important and
central one because it allows the faithful conservation of ge-
netic information, and therefore the propagation of successful
genotypes. Other mechanisms involved in biological inheri-
tance play a role by contributing to this central phenomenon.
Cultural propagation, on the other hand, is achieved through
many different and independent mechanisms, none of which is
central and none of which is a robust replication mechanism.
Limits of Imitation
Imitation is often presented as the main mechanism of cultural
propagation, but, in fact, the notion is stretched to cover a wide
variety of quite different processes. It might seem reasonable,
for instance, to suggest that both the sounds of words and their
written shapes are culturally transmitted through imitation. In
fact, learning the sounds of words is a universal, automatic,
highly canalized process taking place in early childhood and
needing no explicit teaching. Learning the shapes of written
words, on the other hand, is specific to certain cultural groups,
takes place in later childhood or in adulthood and involves
deliberate effort on the part of learners and teachers. Speech
sounds and writing are cultural items that involve different
cognitive and social mechanisms, and also different brain ar-
eas (Mazoyer et al. 1993; Cohen and Dehaene 2004). As for
the meaning of words, it cannot be perceived and therefore
cannot be imitated. Its cultural transmission involves yet other
cognitive and social mechanisms (Bloom 2000).
The Argument from Macrostability
Still, the diversity of mechanisms involved in cultural propa-
gation would not matter too much to the Darwinian program
July 23, 2005; accepted August 2, 2005
c 2006 Konrad Lorenz Institute for Evolution and Cognition Research

if, on the whole, these mechanisms produced the kind of
heritability that makes selection possible. Darwin had good
commonsense reasons to believe in the robustness of biolog-
ical inheritance and this allowed him to develop his theory
without any knowledge of the mechanisms that secured in-
heritance. Is not the reality of cultural inheritance as blatant
as that of biological inheritance? Is not the very existence of
culture—that is, of relatively stable representations, practices,
and artifacts distributed across generations throughout a so-
cial group—sufficient evidence of the heritability of cultural
items? If the microprocesses of cultural propagation failed to
exhibit sufficient fidelity, how could the relative stability evi-
denced by the very existence of culture ever be arrived at? This
argument from observed macrostability to inferred microher-
itability ignores the possibility that microheritability may not
be the only factor capable of securing macrostability.
The various mental and social microprocesses that to-
gether achieve cultural propagation combine preservative and
constructive functions in different degrees. (We talk of “preser-
vative” rather than “reproductive” or “replicative” function,
because, unlike genes, cultural material does not self-
reproduce; rather, it serves as input to intra- and interindividual
processes that tend to preserve their informational content in
their output.) If, instead of postulating that the preservative
function of memory, imitation, and communication must be
effective enough to explain cultural macrostability, one pays
attention to the vast psychological and sociological literature
on these processes, then it becomes clear that high fidelity is
the exception rather than the rule.
Constructive Propagation
Outputs of individual memory, imitation, and communication
processes are, quite generally, not copies but transformations
of the inputs. This is due in part to the imperfection of these
mechanisms: some information is just lost in the process of
transmission. It is also due to the very function of these mech-
anisms, which is never purely preservative. Even when they
are faithful, the microprocesses of cultural propagation are in
good part constructive rather than replicative (Sperber 1996,
2000). The preservation of information is a means toward the
end of providing individuals with information adapted to their
mental and motor schemas and to the pursuit of their goals.
The output of the mechanisms involved is only in part based on
the input information to be remembered or transmitted; it also
draws on general and contextual information and is shaped by
the constructive functions of these mechanisms.
The fact that processes of cultural propagation are partly
preservative and partly constructive affects the ways in which
cultural stability itself can be explained. If the processes in-
volved were just preservative, an occasional error of replica-
tion (akin to a mutation) would be preserved in further errorless
Biological Theory 1(1) 2006
Dan Sperber and Nicolas Claidière
replications: it would become the model. If such copying errors
were frequent, heritability might be too low—or, equivalently,
the mutation rate might be too high—for selection to be effec-
tive, or for anything to propagate at all. Given that the processes
involved in cultural propagation are partly constructive, fail-
ures to faithfully reproduce the inputs of these processes are
not, properly speaking, “copy errors.” Rather than resulting in
random drift, these constructive transformations may be bi-
ased toward producing outcomes that are more suited to the
individual’s cognitive and practical abilities and goals.
Here is how the partly constructive character of the pro-
cesses involved in cultural propagation helps explain cultural
stability (alongside other mechanisms such as the decision bi-
ases described by Richerson and Boyd 2005; see Sperber and
Claidière in press for discussion). Many constructive biases
are shared in a population. This may be due to the fact that
they are based on the common psychological makeup of the
species, or to more local historical or ecological factors. Shared
biases may not only statistically compensate the low fidelity of
preservative microprocesses; they may also permit some items
to reach a cultural level of distribution and stability.
Here is a simple example. The region of the continuum
of colors referred to by a given color term, say “red,” does not
have clear boundaries, but it has, for every user, a focal point
which is seen as prototypical red. In learning the meaning
of “red,” a child is not taking the first sample he/she hears
described as “red” as prototypical red, nor is he/she averaging
over all the samples that he/she hears described as “red” in
his/her learning period. Rather, the child’s color perception
system provides him/her with a ready-made focal point for
red on the basis of which he/she interprets the word. The child
may depart from the samples of “red” he/she is presented with,
in the direction of what is a more salient identification, given
his/her perceptual dispositions. Because the color focal points
for red and other basic colors are similar across individuals,
they stabilize rough common meanings for basic color terms in
any given language, and terms with the same focal points
across many languages, in spite of the fact that interlocutors
hardly ever mean exactly the same thing by “red” (Berlin and
Kay 1969; Kay et al. 1997).
Cultural Causality Is Promiscuous
We can now spell out the most profound disanalogy between
biological and cultural evolution. Among the causal factors
determining biological evolution, there is sharp distinction be-
tween, on the one hand, an extraordinarily robust mechanism
of inheritance—replication—without which there would be no
biological evolution at all, and, on the other hand, a variety of
environmental factors (physical factors, inter- and intraspecies
factors, intergene factors, cytoplasmic environment, and so
on) that affect evolution by determining the relative success
21
Why Modeling Cultural Evolution Is Still Such a Challenge
of different genes. The factors determining cultural evolution
cannot be similarly divided into replication mechanisms and
environmental factors. While some mechanisms (of memory,
imitation, and communication) have a primarily preservative
function, they themselves involve constructive processes draw-
ing extensively on contextual information delivered by other
mechanisms (for instance mental mechanisms of perception,
inference, or action planning, and social processes of seduc-
tion, coercion, deception, bargaining, coalition formation, and
so on), the primary function of which is not preservative.
Constructive processes systematically intrude on preservative
processes. Cultural contents are not replicated by one set of
inheritance mechanisms and selected by another, disjoint set
of environmental factors—not by a long shot.
We are well aware of the fact that biological transmis-
sion is much more complex than a few sentences can indicate,
and that it involves a variety of submechanisms that may have
analogues in cultural evolution. To give just one illustration,
consider an orally transmitted folktale. The public telling of
the tale contributes to the formation in the listeners of mental
representations of the tale, and these mental representations
contribute to the production of further public telling by lis-
teners turned tellers. The causal chain involves an alternation
of mental and public events with equally potent causal roles.
The case of RNA retroviruses that propagate through reverse
transcription into DNA might be seen as a biological analogue.
But how helpful would such an analogy be?
Merely adjusting the general model of Darwinian selec-
tion to describe cultural evolution involves idealizing away
properties crucial to the proper explanation of the phenomena.
Looking to biology for fine-grained analogies between biolog-
ical and cultural evolution may be a source of insights but it
should not take precedence over looking to the cognitive and
social sciences for a better understanding of the actual mech-
anisms and processes involved. We agree with Richerson and
Boyd (2005) that Darwin-inspired population thinking pro-
vides the proper approach to the scientific study of cultural
evolution. However, it has yet to be generalized to cases where
the causal forces involved cannot be neatly separated into
22
reproductive mechanisms and environmental factors. Cultural
causality is promiscuous and that is why modeling it is still
such a challenge.
References
Aunger R (2002) The Electric Meme: A New Theory of How We Think. New
York: Free Press.
Berlin B, Kay P (1969) Basic Color Terms: Their Universality and Evolution.
Berkeley: University of California Press.
Bloom P (2000) How Children Learn the Meanings of Words. Cambridge,
MA: MIT Press.
Boyd R, Richerson PJ (1985) Culture and the Evolutionary Process. Chicago:
University of Chicago Press.
Campbell DT (1960) Blind variation and selective retention in creative thought
as in other knowledge processes. Psychological Review 67: 380–400.
Cavalli-Sforza LL, Feldman MW (1981) Cultural Transmission and Evolution.
Princeton: Princeton University Press.
Cohen L, Dehaene S (2004) Specialization within the ventral stream: The case
for the visual word form area. NeuroImage 22: 466–476.
Dawkins R (1976) The Selfish Gene. Oxford: Oxford University Press.
Durham W (1991) Coevolution: Genes, Culture, and Human Diversity. Stan-
ford: Stanford University Press.
Kay P, Berlin B, Maffi L, Merrifield W (1997) Color naming across languages.
In: Color Categories in Thought and Language (Hardin CL, Maffi L, eds),
21–58. Cambridge: Cambridge University Press.
Mazoyer BM, Tzourio N, Frak V, Syrota A, Murayama N, Levrier O, Salamon
G, Dehaene S, Cohen L, Mehler J (1993) The cortical representation of
speech. Journal of Cognitive Neuroscience 5: 467–479.
Mesoudi A, Whiten A, Laland K (2004) Is human cultural evolution Dar-
winian? Evidence reviewed from the perspective of The origin of species.
Evolution 58: 1–11.
Monod J (1970) Le hasard et la nécessité. Paris: Seuil.
Richerson PJ, Boyd R (2005) Not by Genes Alone. Chicago: University of
Chicago Press.
Sperber D (1996) Explaining Culture: A Naturalistic Approach. Oxford:
Blackwell.
Sperber D (2000) An objection to the memetic approach to culture. In: Dar-
winizing Culture (Aunger R, ed), 163–174. Oxford: Oxford University
Press.
Sperber D, Claidière N. (in press) Defining and explaining culture (comments
on Richerson and Boyd, Not by genes alone). Biology and Philosophy.
Sperber D, Hirschfeld L (1999) Evolution, cognition and culture. In: MIT
Encyclopedia of the Cognitive Sciences (Keil F, Wilson R, eds), cxi–
cxxxii. Cambridge, MA: MIT Press.
Biological Theory 1(1) 2006